A Multi-Dimensional Evaluation of Global Ocean salinity in CMIP6 Historical simulations

A realistic representation of ocean salinity is essential for understanding large-scale ocean circulation, water mass transformation, and the global hydrological cycle. This study thoroughly assesses CMIP6 historical simulations by comparing ocean salinity outputs from climate models with EN4 datasets, investigating horizontal and vertical spatial patterns as well as temporal variability. Our analysis shows that while CMIP6 models display a strong spatial correlation with observations across multiple ocean basins, significant discrepancies arise in the temporal trends of salinity variations. Specifically, the models often hard to capture the magnitude and persistence of long-term salinity trends observed in the EN4 dataset. The evaluation also emphasizes vertical salinity profiles, where maximum and minimum salinity values serve as simplified indicators for structural assessment. To ensure consistent comparisons, model-simulated salinity extrema are standardized by aligning their values with the corresponding depths in the observational data. These results highlight both the strengths and limitations of CMIP6 models in representing key oceanographic features, such as the depth and magnitude of salinity extrema, mixed layer depth, and regional salinity gradients. By identifying areas of agreement and divergence between models and observations, this study provides valuable insights for improving the physical realism and predictive accuracy of historical climate simulations, ultimately guiding future model development.